Reduced-Order Microstructure-Sensitive Models for Damage Initiation in Two-Phase Composites

• Published in: Integrating materials and manufacturing innovation Vol. 7; no. 3; pp. 97 - 115
• Main Authors: Montes de Oca Zapiain, David, Popova, Evdokia, Abdeljawad, Fadi, Foulk, James W., Kalidindi, Surya R., Lim, Hojun
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.09.2018; Springer Nature B.V
• Subjects: Characterization and Evaluation of Materials; Chemistry and Materials Science; Composite materials; Correlation; Crack initiation; Damage; Extreme values; Fracture mechanics; Machine learning; Manufacturing; Materials research; Materials Science; Metallic Materials; Microstructure; Nanotechnology; Properties (attributes); Property damage; Reduced order models; Structural damage; Structural Materials; Studies; Surfaces and Interfaces; Technical Article; Thin Films; Materials Knowledge System; 2-point statistics; Damage; Extreme value distributions; PCA
• ISSN: 2193-9764; 2193-9772
• DOI: 10.1007/s40192-018-0112-0

Local features of the internal structure or the microstructure dominate the overall performance of materials. An open problem in materials design with enhanced properties is to accurately identify and quantify salient features of the microstructure and understand its correlation with the material’s performance. This task is exacerbated when dealing with failure related properties that show strong correlations to higher-order details of the material microstructure. This paper presents a novel data-driven framework for quantitatively determining the highly complex correlations that exist between the higher-order details of the material microstructure and its failure-related properties, specifically its damage initiation properties. The enclosed work will address this challenge by significantly extending the Materials Knowledge Systems (MKS) framework and by leveraging concepts in extreme value distributions and machine learning. The developed framework was capable of successfully sorting nine different classes of synthetically generated two-phase microstructures for their sensitivity to damage initiation. The framework and approaches presented here open new research avenues for studying the microstructure-sensitive damage initiation properties associated with heterogeneous materials, and pave the way forward for practical multiscale materials design.